In a quasi-experimental design, targeted youth were assigned to one of two levels of treatment (classroom sessions or a combination of classroom sessions and structured after-school activities) or to a no-treatment control group. To ascertain gang membership following the intervention, the youths' names were compared with gang membership rosters obtained from informants. Four members of the control group and one youth receiving the more intensive intervention were subsequently identified as gang members. Implications of the study for public policy are considered, methodological difficulties in conducting research in this area are discussed, and future research directions are suggested. (Author abstract)